Analysis of transmission spatio-temporal pattern of atmospheric heavy pollution events based on spatiotemporal data mining

In order to explore the spatiotemporal pattern of regional air pollution transmission and diffusion and to support the regional joint prevention and emergency control of atmospheric environment pollution, a new spatiotemporal data mining approach was proposed based on air-quality ground observation data. The algorithms were built to identify the regional transmission paths and intensity of heavily polluted air masses. The case study on PM_2.5 heavy pollution events in Beijing-Tianjin-Hebei region during January-March and October-December, 2021 was conducted to verify the algorithms. The results showed that there were 17 regional heavy pollution events during this period in Beijing-Tianjin-Hebei region. There were 3, 7, and 7 long (>48 h), medium-long (24-48 h), and short (<24 h) pollution events, respectively. All long pollution events occurred in spring. Their pollution intensity was the highest among the three event types, and the polluted area and pollution transmission area covered the entire study area. The medium-long and short pollution events occurred in spring and winter. Their pollution intensity was lower than that of long events. Medium-long events' polluted area coverage (>80%) was higher than that of short events (<63%). There were seasonal differences in the area covered by pollution transmission for medium-long pollution events. In terms of the transportation intensity coefficient of heavy pollution transmission in Beijing-Tianjin-Hebei region generally conformed to the law that local pollution intensity coefficient (0.32-1.00) was the highest, followed by intra-city pollution transmission (0.01-0.95), and inter-city pollution (0.00-0.28). Among them, the inter-city transmission intensity of Xingtai was greater than its intra-city transmission, and the impact of Hengshui on surrounding cities was lower than the average level.